JPS60176185A - Quantizing method of hand-written character and graphic - Google Patents

Quantizing method of hand-written character and graphic

Info

Publication number
JPS60176185A
JPS60176185A JP59030279A JP3027984A JPS60176185A JP S60176185 A JPS60176185 A JP S60176185A JP 59030279 A JP59030279 A JP 59030279A JP 3027984 A JP3027984 A JP 3027984A JP S60176185 A JPS60176185 A JP S60176185A
Authority
JP
Japan
Prior art keywords
temporary
line segment
segment
information
direction code
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP59030279A
Other languages
Japanese (ja)
Inventor
Takanori Yokoyama
孝典 横山
Soshiro Kuzunuki
壮四郎 葛貫
Hiroshi Shiyoujima
博 正嶋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hitachi Ltd
Original Assignee
Hitachi Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hitachi Ltd filed Critical Hitachi Ltd
Priority to JP59030279A priority Critical patent/JPS60176185A/en
Publication of JPS60176185A publication Critical patent/JPS60176185A/en
Pending legal-status Critical Current

Links

Landscapes

  • Image Processing (AREA)
  • Character Discrimination (AREA)

Abstract

PURPOSE:To enable quantization free from a character fluctuation and an inclination by extracting a temporary segment from inputted coordinate information and by considering the synthesization of adjacent temporary segments as one temporary segment if the difference in the temporary directional code of the adjacent temporary segments. CONSTITUTION:Coordinate information (a) obtained from a coordinate input device 1 is inputted to a temporary segmenting processing part 2, a candidate considered to be the segment is extracted and set to temporary segment information (b), which is inputted to a temporary directional code generating part 3. In accordance with the direction of the temporary segment, temporary directional code information (c) in 32 directions is generated and inputted to a segment synthesization processing part 4, which synthesizes the segment if necessary by utilizing the temporary directional code information (c) and outputs the final segment information (d). A directional code generating part 5 outputs directional code information (e) in eight objective directions in accordance with the direction of the segment. Provided that when directional codes of adjacent two segments are equal, they are synthesized in order to be one line segment, and one directional code shall be given.

Description

【発明の詳細な説明】 〔発明の利用分野〕 本発明は、図形あるいは文字を植成する線分の方向を特
徴量として使用するパターン認識装置に係り、特に、手
書き図形あるいは文字に特有なふらつきや傾きの影響を
受けずに特徴を抽出するのに好適な量子化方法に関する
[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a pattern recognition device that uses the direction of a line segment in which a figure or a character is placed as a feature quantity, and in particular, it relates to a pattern recognition device that uses the direction of a line segment in which a figure or a character is placed as a feature quantity, and in particular, The present invention relates to a quantization method suitable for extracting features without being affected by gradients or slopes.

〔発明の背景〕[Background of the invention]

従来、図形や文字などのパターン認識装置は。 Conventionally, pattern recognition devices such as figures and characters.

同形や文字を構成する線分の方向を特徴量、あるいは、
その一部とする手法が数多く用いられている。しかし、
この方法により手書き図形あるいは文字を認識する場合
、手書きの特徴であるふらつきや傾きの影響により線分
の数やその方向を的確に抽出することができず、認識率
を低下させるという問題があった。特に、量子化を行な
う際の方向範囲の境界(閾値)付近でのふらつきの影響
を取り除くことが重要である。
The direction of the line segments that make up the isomorphism or character is used as a feature quantity, or
Many methods are used as part of this. but,
When recognizing handwritten figures or characters using this method, there was a problem in that the number and direction of line segments could not be accurately extracted due to the effects of wobbling and inclination, which are characteristics of handwriting, and the recognition rate decreased. . In particular, it is important to remove the influence of fluctuation near the boundary (threshold) of the directional range when performing quantization.

この問題を解決する方法として、日経エレクトロニクス
、1973年5月7日号、P46〜47に示された例や
昭和56年度電子通信学会総合全国大会予稿集No13
76に開示された例などがある。
As a method to solve this problem, the example shown in Nikkei Electronics, May 7, 1973 issue, pages 46-47, and the Proceedings of the 1986 Institute of Electronics and Communication Engineers National Conference No. 13
Examples include the one disclosed in No. 76.

前者は、それ以前に量子化された部分の最終の量子化方
向の許容範囲を他の方向より拡大させることによって、
ふらつきの影響を減少させている。
The former is by expanding the tolerance range of the final quantization direction of the previously quantized part compared to other directions.
Reduces the effects of wandering.

しかし、この方法では許容範囲が固定化されているため
に、量子化範囲の境界付近でのふらつきの影響を完全に
なくすことは不可能であるという欠点があった。
However, this method has a drawback in that it is impossible to completely eliminate the influence of fluctuations near the boundaries of the quantization range because the tolerance range is fixed.

一方、後者は、量子化する際の境界を変えて量子化を二
回行ない、いずれかの方向となったものを一本の線分と
見なすことによって、境界付近での線分のふらつきの影
響を減少させている。しかし、この方法は二回の量子化
の結果を合成するという複雑な処理を行なうために、ハ
ードウェア化が麗しいことや、入力線分の絶対方向の影
響を完全には無くシていないという欠点があった。
On the other hand, in the latter case, the quantization is performed twice by changing the boundary during quantization, and by considering either direction as a single line segment, the influence of the fluctuation of the line segment near the boundary is decreasing. However, since this method performs a complex process of composing the results of two quantizations, it is elegant to implement in hardware, and the disadvantage is that it does not completely eliminate the influence of the absolute direction of the input line segment. was there.

〔発明の目的〕[Purpose of the invention]

本発明の目的は、こうした欠点を克服するため、入力さ
れた図形あるいは文字のふらつきや傾きの影響を受けな
い量子化方法を提供することにある。
SUMMARY OF THE INVENTION In order to overcome these drawbacks, it is an object of the present invention to provide a quantization method that is not affected by the fluctuation or inclination of input figures or characters.

〔発明の概要〕[Summary of the invention]

本発明は、入力された座標情報から、代表的な点列を結
んだ仮線分を抽出し、この仮線分に、特徴量として用い
る方向コードより細分化した仮方向コードを与え、次に
、隣接した二つの仮線分の仮方向コードの差が所定値以
下であればそれらは同一の線分に含まれるとして、それ
らを合成したものを改めて一つの仮線分とし、所定値以
上であれば、別の線分と見なす処理を逐次行なうことに
よって、最適な線分化を行ない、その上で特徴量とする
方向コードを発生することによって、入力された図形あ
るいは文字の傾きによらずに、手書き特有のふらつきの
影響を受ずに、量子化を行なうものである。
The present invention extracts a temporary line segment connecting representative point sequences from input coordinate information, gives this temporary line segment a temporary direction code that is subdivided from the direction code used as a feature, and then , if the difference in the temporary direction codes of two adjacent temporary line segments is less than a predetermined value, they are considered to be included in the same line segment, and the composite of them is considered to be a single temporary line segment, and if the difference is more than a predetermined value If there is, it is possible to perform optimal line segmentation by sequentially performing processing to consider it as a different line segment, and then generate a direction code as a feature, regardless of the slope of the input figure or character. , which performs quantization without being affected by the fluctuations peculiar to handwriting.

[発明の実施例〕 以下、本発明の一実施例を第1図ないし第6図を用いて
詳細に説明する。
[Embodiment of the Invention] Hereinafter, an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 6.

第1図は本発明の一実施例のブロック図である。FIG. 1 is a block diagram of one embodiment of the present invention.

この実施例では、図形あるいは文字を構成する成分の方
向に従って、第2図に示すような8種の方向コードを発
生し、これを特徴量として認識を行なうものである。座
標入力装置1によって得られた座標情報aは、仮線分化
処理部2に入力される。
In this embodiment, eight types of direction codes as shown in FIG. 2 are generated according to the direction of the components constituting a figure or character, and these are recognized as feature quantities. Coordinate information a obtained by the coordinate input device 1 is input to the temporary line differentiation processing section 2.

仮線分化処理部2は、入力された座標情報から線分と思
われる候補を抽出するもので、これには種種の方式があ
るが、例えば、入力座標を結んだ曲線の角度変化を利用
する方法がある。これは、角度変化の分布を調べ、角度
変化が所定値以上となる点が線分の境界であるとするも
のである。すなわち、角度変化が所定値以上となる二点
間を−・つの線分とする方法である。いずれの方法を用
いるとしても、この仮線分化処理部2では、入力された
座標情報aが手書図形あるいは文字のものであるため、
本来一本の線分と見なされるべきものが線分のふらつき
によってそれ以上の数の線分に分割されてしまうことが
多い。
The temporary line differentiation processing unit 2 extracts candidates that are considered to be line segments from the input coordinate information.There are various methods for this, but for example, it uses the angle change of a curve connecting the input coordinates. There is a way. This is done by examining the distribution of angular changes and determining that the points where the angular changes are equal to or greater than a predetermined value are the boundaries of line segments. That is, this is a method in which the distance between two points where the angle change is equal to or greater than a predetermined value is defined as -. line segments. Regardless of which method is used, in this temporary line differentiation processing unit 2, since the input coordinate information a is of a handwritten figure or a character,
What should originally be considered a single line segment is often divided into more line segments due to line segment fluctuation.

仮線分化処理部2によって得られた仮線分情報すは、仮
方向コード発生部3に入力され、仮線分の方向に従って
第3図に示すような、O1〜32の32方向の仮方向コ
ードを発生する。この仮方向コード情報Cは線分合成処
理部4に入力される。
The temporary line segment information obtained by the temporary line segmentation processing unit 2 is input to the temporary direction code generation unit 3, and temporary directions are generated in 32 directions O1 to O32 as shown in FIG. 3 according to the direction of the temporary line segment. Generate code. This temporary direction code information C is input to the line segment synthesis processing section 4.

線分合成処理部4では、板方向コード清面Cを用いて必
要に応じて線分の合成を行ない、最終的な線分情報dを
出力する。方向コード発生部5は線分情報dから、線分
の方向に従って第2図に示すような目的とする8方向の
方向コード情報Cを出力する。ただし、隣り合う二線分
の方向コードが等しい場合には、合成して一つの線分と
し、一つの方向コードを与えるものとする。
The line segment synthesis processing section 4 synthesizes line segments as necessary using the board direction code clear side C, and outputs final line segment information d. From the line segment information d, the direction code generating section 5 outputs direction code information C for eight target directions as shown in FIG. 2 according to the direction of the line segment. However, if the direction codes of two adjacent line segments are the same, they are combined into one line segment and given one direction code.

方向コードマツチング処理部6では1手書き図形あるい
は文字の方向コード情報eと、あらかじめ辞書7に登録
しておいた方向コード情報fを比較し、最も近いと判断
される図形、あるいは、文字を選出し゛、その図形ある
いは文字情報gを図形・文字表示装置8に出力し、図形
文字表示装置8でLFA識された図形あるいは文字を表
示する。こ九ら6,7.8については多くの例が知られ
ているので、ここでは詳細な説明は省略する。
The direction code matching processing unit 6 compares the direction code information e of one handwritten figure or character with the direction code information f registered in advance in the dictionary 7, and selects the figure or character that is judged to be the closest. Then, the graphic or character information g is output to the graphic/character display device 8, and the graphic or character information recognized by LFA is displayed on the graphic/character display device 8. Since many examples of these 6, 7, and 8 are known, detailed explanations will be omitted here.

ここで、本発明の特徴である線分合成処理部4の処理内
容について以下、詳細に説明する。第4図は線分合成処
理を示した流れ図である。まず、最初の仮線分とその接
続されている二番目の仮線分との板方向コードの差をめ
る(ステップ402)。ただし、この差は、この方向コ
ード値そのものの差ではなく、角度の差どし、たとえば
、01と32の差は1とする。そして、この差が3以下
であれば第一の仮線分と第二の仮線分は同一の線分に含
まれているものと見なし、この二つを合成したものを改
めて一つの仮線分し、この台分した仮線分の板方向コー
ドをめる(ステップ406)。そして、この合成した仮
線分と、その次の仮線分、すなわち、三番目の仮線分と
の間で前述と同様な板方向コードの比較を行なう。次に
、方向コードの差が4以上の場合であるが、この場合に
は、二つの仮線分はそれぞれ別の線分と見なし、第一の
仮線分は一つの線分であると判断しくステップ404)
、改めて、第二の仮線分とその次の仮線分の間で板方向
コードの比較を行なう。
Hereinafter, the processing contents of the line segment synthesis processing section 4, which is a feature of the present invention, will be explained in detail. FIG. 4 is a flowchart showing the line segment combining process. First, the difference in sheet direction code between the first temporary line segment and the second temporary line segment to which it is connected is calculated (step 402). However, this difference is not the difference in the direction code value itself, but the difference in angle, for example, the difference between 01 and 32 is assumed to be 1. If this difference is 3 or less, the first temporary line segment and the second temporary line segment are considered to be included in the same line segment, and the composite of these two is reused as one temporary line. Then, the board direction code for the temporary line segment is entered (step 406). Then, the same sheet direction code as described above is compared between this combined temporary line segment and the next temporary line segment, that is, the third temporary line segment. Next, when the difference in direction codes is 4 or more, in this case, the two temporary line segments are considered to be separate line segments, and the first temporary line segment is determined to be one line segment. Step 404)
, the board direction codes are again compared between the second temporary line segment and the next temporary line segment.

以下同様の処理を続けることによって、真の線分の決定
を行なう。
By continuing similar processing, the true line segment is determined.

以−Hに述べた処理を行なった例を第5図に示す。FIG. 5 shows an example in which the processing described in section H below is performed.

(a)は仮線分化処理を終了した手書き曲線であり、Δ
1〜A4の四つの仮線分で構成され、それぞれの板方向
コードは、02,03,02.t。
(a) is a handwritten curve that has undergone the temporary line differentiation process, and Δ
It is composed of four temporary line segments from 1 to A4, and the board direction codes are 02, 03, 02. t.

となっている。まず仮線分A1とA2の板方向コードを
比較するど、その差は】であるから、これらは同一の線
分に含まれていると見なし、これを合成したものを改め
て一つの仮線分Δ5とする。
It becomes. First, compare the plate direction codes of temporary line segments A1 and A2, and the difference is ], so these are considered to be included in the same line segment, and the composite of these is re-created as one temporary line segment. Let it be Δ5.

こJl、が(b)である。仮線分A5の板方向コードは
03となる。次に、仮線分A5とA3の方向コートを比
較すると、その差は1であるから、仮線分A5.A3を
合成し、仮線公人6を得る。八6の板方向コードは02
である。これが(C)である。次に、仮線分A6とA4
を比較するとその板方向コード差は8であるから、A6
は独立した線分であると決定する。また、A4は最後の
仮線分であるから、これも一つの線分とする。このよう
にして、(、)のような四つの仮線分の組み合わせは、
 (c)のような二つの線分と見なされる。
This Jl is (b). The board direction code of the temporary line segment A5 is 03. Next, when comparing the direction coats of temporary line segments A5 and A3, the difference is 1, so temporary line segment A5. Synthesize A3 and obtain the temporary public figure 6. The board direction code for 86 is 02
It is. This is (C). Next, temporary line segments A6 and A4
When compared, the board direction code difference is 8, so A6
are independent line segments. Furthermore, since A4 is the last temporary line segment, it is also treated as one line segment. In this way, the combination of four temporary line segments like (,) is
It is considered as two line segments as shown in (c).

なお、線分情報は、第6図に示すように、方向コード発
生部に送られ、DI、D3という、目的とする方向コー
ドを出力し、方向コードマツチング処理部6へ送られる
As shown in FIG. 6, the line segment information is sent to the direction code generation section, which outputs the desired direction codes DI and D3, and sent to the direction code matching processing section 6.

次に本発明の実施例の効果について述べる。Next, the effects of the embodiments of the present invention will be described.

第一に、仮線分化を行なった後に、必要によって線分合
成処理を行なって最終的な線分を決定するという二段階
方式をとっているため、ふらつきのある長い線分でも複
数線分に分割されず、短い線分も見落とされないという
効果がある。
First, it uses a two-step method in which after performing temporary line segmentation, line segment compositing processing is performed as necessary to determine the final line segment, so even long line segments with fluctuations can be combined into multiple line segments. It has the effect that it is not divided and short line segments are not overlooked.

線分化を一段階で行なう方法では、/M′iい線分とふ
らつきの違いを検出しにくく一般に、ふらつきのある長
い線分を的確に抽出しようとすると、短い線分を見落と
し、逆に、短い線分をも見落しせずに抽出しようとする
と、ふらつきのある長い線分を複数線分に分割してしま
うという欠点があった。本実施例では線分化を二段階に
行なうことによって、この欠点製克服した。
In a method that performs line segmentation in one step, it is difficult to detect the difference between /M'i long line segments and wobbling.In general, when trying to accurately extract long line segments with wobbling, short line segments are overlooked, If an attempt is made to extract short line segments without overlooking them, the problem is that long line segments with fluctuations are divided into multiple line segments. In this example, this drawback was overcome by performing line segmentation in two stages.

第二に、手書き図形、あるいは、文字を構成する線分が
皿子化の境界付近に存在する場合でも手書き特有のふら
つきの影響を受けずに旦子化を行なえるという効果があ
る。
Second, even if a handwritten figure or a line segment constituting a character exists near the border of slanting, it can be slanted without being affected by the unsteadiness peculiar to handwriting.

第6図に示すような、A1−A4およびB1〜B4とい
う仮線分列について、以下その効果を述べる。最終的に
得る方向コードは第2図に示すような8方向である。第
6図中、TI、]”2.1’3は方向コード化の境界で
ある。仮線分列13]、〜B4の板方向コードは、08
−09−08−16であり、線分合成処理を行なうとB
 1〜133の仮線分例は一つの線分となり、最終的な
方向コードは、同一方向コードの連続した線分は一つの
線分とみなしD3−D5となる。この仮線分列について
は、一段階で最終コードを発生ずる方法でも同様の結果
が得られる。一方、仮線分列Al−A4の板方向コード
は、02−03−02−] Oであり1本発明で示すよ
うな線分合成処理を行なえば。
The effects of the temporary line segment sequences A1-A4 and B1-B4 as shown in FIG. 6 will be described below. The direction codes finally obtained are eight directions as shown in FIG. In FIG. 6, TI, ]"2.1'3 is the boundary of direction coding. The board direction code of the temporary line segment sequence 13], ~B4 is 08
-09-08-16, and when line segment synthesis processing is performed, B
The tentative line segment examples 1 to 133 become one line segment, and the final direction code becomes D3-D5 since continuous line segments with the same direction code are regarded as one line segment. Regarding this temporary line segment sequence, a similar result can be obtained by a method in which the final code is generated in one step. On the other hand, the board direction code of the temporary line segment sequence Al-A4 is 02-03-02-]O, and if line segment synthesis processing as shown in the present invention is performed.

第5図に示すように、二つの線分となり、最終的な方向
コードは、DI−D3となる。ところが、この仮線分列
について、一段階で最終コードを発生するとD 1−D
2−D 1−03となり、仮線分列A1〜Δ4と81〜
B4は方向が60″変化しただけで、全く同じ形をして
いるのにも関らず、線分の数が違っている。すなわち、
境界付近のA1−A4の仮線分列はふらつきの影響を受
けていることになる。一方、本実施例では、隣り合う二
つの仮線分の仮方向コードの差によって最終的な線分の
決定を行なっているため、入力線分の絶対的な方向の影
響(X子化誤差)は全く存在しない。
As shown in FIG. 5, there are two line segments, and the final direction code is DI-D3. However, when the final code is generated in one step for this temporary line segment sequence, D 1-D
2-D 1-03, and the temporary line segment sequences A1~Δ4 and 81~
Although B4 has the exact same shape with only a 60" change in direction, the number of line segments is different. In other words,
The temporary line segment sequence A1-A4 near the boundary is affected by the fluctuation. On the other hand, in this embodiment, since the final line segment is determined based on the difference between the temporary direction codes of two adjacent temporary line segments, the influence of the absolute direction of the input line segment (X-conversion error) does not exist at all.

本発明の実施例では、仮方向コードを32方向、最終的
な方向コードを8方向としているが、これらの方向数は
場合に応じて適当に変えてよい。また、仮線分を合成す
るかどうかを決定する仮方向コード差の閾値(第4図、
402中のd)も場合に応じて適当に変えてよい。さら
に、仮線分化処理を行なわずに、入力座標を一定間隔で
サンプリングし、そのサンプリング点間を一つの仮線分
と見なして、その後の処理を同様に行なうことも可能で
ある。
In the embodiment of the present invention, the temporary direction code is 32 directions and the final direction code is 8 directions, but the number of directions may be changed as appropriate depending on the situation. In addition, the threshold value of the temporary direction code difference (Fig. 4,
d) in 402 may also be changed as appropriate. Furthermore, it is also possible to sample the input coordinates at regular intervals without performing the temporary line segmentation process, treat the sampling points as one temporary line segment, and perform subsequent processing in the same way.

本発明の実施例は1Mt終的に方向コードマツチング処
理を行ない、図形、あるいは、文字の認識を行なってい
るが、線分化の結果を表示するだけでも1図形清書力式
として使用可能である。
The embodiment of the present invention performs 1Mt final direction code matching processing to recognize figures or characters, but it can also be used as a single figure formatting function just by displaying the result of line segmentation. .

〔発明の効果〕〔Effect of the invention〕

本発明によれば、入力された手書き図形、あるいは、文
字の傾きによらず、言い換えれば、入力された手IFI
IIA分の絶対的な方向にかかわらず、手書き特有のふ
らつきの影響を受けずに、線分の量子化ができるので、
手書き認識装置の認識率を上げることができる。
According to the present invention, regardless of the input handwritten figure or the inclination of the character, in other words, the input hand IFI
Regardless of the absolute direction of the IIA component, line segments can be quantized without being affected by the fluctuations peculiar to handwriting.
The recognition rate of the handwriting recognition device can be increased.

【図面の簡単な説明】[Brief explanation of the drawing]

第1図は本発明の一実施例の全体ブロック図、第2図、
第3図は本発明の一実施例の方向コード図、第4図は本
発明の一実施例の流れ図、第5図。 第6図は本発明の一実施例の処理の説明図である。 1・・・座標入力装置、2・・・仮線分化処理部、3・
・・仮方向コード発生部、4・・・線分合成処理部、5
・・・方向コード発生部、6・・・方向コードマツチン
グ処理部 2 日 3 7 第3目 3 第 4 口 第 5 日 第 6 凹
FIG. 1 is an overall block diagram of an embodiment of the present invention, FIG.
FIG. 3 is a direction code diagram of an embodiment of the present invention, FIG. 4 is a flowchart of an embodiment of the present invention, and FIG. 5 is a diagram. FIG. 6 is an explanatory diagram of processing according to an embodiment of the present invention. 1... Coordinate input device, 2... Temporary line differentiation processing unit, 3.
... Temporary direction code generation section, 4... Line segment synthesis processing section, 5
... Direction code generation section, 6 ... Direction code matching processing section 2nd day 3 7 3rd eye 3rd mouth 5th day 6th concave

Claims (1)

【特許請求の範囲】 1、線分の方向を特徴量として使用する手書き文字・図
形の確認方法において。 入力された座標情報から代表的な点列を抽出し、この点
列を結んだ仮の線分に対し、特徴量として使用する方向
コードによって細分化さ九た仮の方向コードを与え、 前記仮の方向コードを前回の仮の線分の方向コードと比
較しで、このコードの差が所定の値以下であれば二つの
線力は同一直線上にあると判定して、前記二つのコード
を合成したものを改めて一つの仮の線分どし、 所定の値より人きけJしば別の仮の線分とする処理を逐
次行なうことにより、最終線分を得て前記特徴量として
使用する方向コードを与えて量子化することを特徴とす
る手書き文字図形の量子化方法。
[Claims] 1. In a method for checking handwritten characters and figures using the direction of a line segment as a feature quantity. A representative point sequence is extracted from the input coordinate information, a temporary line segment connecting this point sequence is subdivided by the direction code used as a feature quantity, and a temporary direction code is given to the temporary line segment. The direction code of is compared with the direction code of the previous temporary line segment, and if the difference between the codes is less than a predetermined value, it is determined that the two line forces are on the same straight line, and the two codes are By sequentially processing the synthesized line segment to create another temporary line segment based on a predetermined value, the final line segment is obtained and used as the feature amount. A method for quantizing handwritten character figures, characterized by quantizing them by giving a direction code.
JP59030279A 1984-02-22 1984-02-22 Quantizing method of hand-written character and graphic Pending JPS60176185A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP59030279A JPS60176185A (en) 1984-02-22 1984-02-22 Quantizing method of hand-written character and graphic

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP59030279A JPS60176185A (en) 1984-02-22 1984-02-22 Quantizing method of hand-written character and graphic

Publications (1)

Publication Number Publication Date
JPS60176185A true JPS60176185A (en) 1985-09-10

Family

ID=12299272

Family Applications (1)

Application Number Title Priority Date Filing Date
JP59030279A Pending JPS60176185A (en) 1984-02-22 1984-02-22 Quantizing method of hand-written character and graphic

Country Status (1)

Country Link
JP (1) JPS60176185A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0488265A (en) * 1990-07-31 1992-03-23 Toyota Motor Corp Metal gasket
WO2002037831A1 (en) * 2000-10-30 2002-05-10 Matsushita Electric Industrial Co., Ltd. Information transmitting/receiving system, information transmitting/receiving method, and handwritten information compressing method used for them

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0488265A (en) * 1990-07-31 1992-03-23 Toyota Motor Corp Metal gasket
WO2002037831A1 (en) * 2000-10-30 2002-05-10 Matsushita Electric Industrial Co., Ltd. Information transmitting/receiving system, information transmitting/receiving method, and handwritten information compressing method used for them

Similar Documents

Publication Publication Date Title
JP2022501663A (en) Equipment and methods for image processing, as well as systems for training neural networks
Sezgin et al. A new dichotomization technique to multilevel thresholding devoted to inspection applications
CN103810473B (en) A kind of target identification method of human object based on HMM
JPH10240943A (en) Test signature inspecting method
JPWO2020062846A5 (en)
Hallale et al. Twelve directional feature extraction for handwritten English character recognition
CN103455816B (en) Stroke width extraction method and device and character recognition method and system
KR910001525B1 (en) Method of curve approximation
US5937093A (en) Pattern recognition methods and apparatus to reduce processing time by groupnise comparison of features and coordinated dynamic elimination of candidates
JPS60176185A (en) Quantizing method of hand-written character and graphic
US7113636B2 (en) Method and computer program product for generating training data for a new class in a pattern recognition classifier
Omarov et al. Machine learning based pattern recognition and classification framework development
CN113344110B (en) Fuzzy image classification method based on super-resolution reconstruction
CN111310548A (en) Method for recognizing stroke type in online handwriting
JPH034953B2 (en)
CN111680683A (en) ROI parameter acquisition method and system
Wai et al. Offline signature verification system using neural network
Venkatesan et al. Advanced classification using genetic algorithm and image segmentation for Improved FD
Yahyatabar et al. Online signature verification: a robust approach for persian signatures
Ahmed et al. English Alphabets Recognition using Hand Gesture
US20110091111A1 (en) Multilevel bit-mapped image analysis method
Xu et al. Collaborative Research on Mouth Shape and Lyrics in Singing Practice Based on Image Processing
CN116246316A (en) Face recognition reinforcing method
JPH06243295A (en) Method and device for generating feature data of character
JP2002541597A (en) Coding and recognition of handwriting